Pooled cancer cell screening identifies genomic determinants of nanoparticle delivery to cancer cells: sub analysis of central nervous system tumor cell lines

Abstract Background There is great interest in utilizing nanoparticles to improve drug delivery to central nervous system (CNS) tumors, but the biologic features underlying successful delivery at the tumor site are not known. METHODS We developed a pooled screening assay to investigate cellular features predictive of nanoparticle delivery and screened 35 fluorescent nanoparticle formulations against 488 pooled cancer cell lines with DNA barcodes. The nanoparticle library was comprised of nonlethal liposomal or polymeric cores with varied surface functionality, including natural and synthetic polymers. Cells were profiled using fluorescence-activated cell sorting and barcodes sequenced and deconvolved to generate an association score for each nanoparticle-cell line pair. RESULTS Of 488 cancer cell lines, 22 cancer cell lineages (tissues of origin) were included. There were 35 cell lines derived from CNS tumors; of these 22 were classified as astrocytoma or glioblastoma, one as oligodendroglioma, and 2 as medulloblastoma. Of CNS tumor cell lines, 40% were from female patients and average age at cell line derivation was 52 years; TP53 hotspot mutations were detected in 25/35 lines, PTEN hotspot mutations in 15/25 lines. Using unsupervised hierarchical clustering by nanoparticle association profiles, CNS tumor lines did not cluster together by lineage, subtype, or mutation status but rather by baseline expression of genes involved in nanocarrier trafficking. Nanoparticle association profiles were heterogeneous among the CNS tumor lines for most formulations, though CNS tumor lines had high association with polystyrene formulations with respect to other lineages. Select CNS tumor lines had high uptake of liposomes coated with the natural polymers fucoidan, alginate, hyaluronic acid and chondroitin sulfate; these trends were further probed in a non-pooled screen of 5 additional cell lines. Conclusions Uptake of liposomal and polymeric nanoparticles is heterogeneous across CNS tumor lines, but can be predicted using baseline gene expression.